Gas pipeline calculations and pressure control 2

[engr2GW]

Hello,
In gas pipelines, using Weymouth, Panhandle, etc. equation depending on which is more suitable given the conditions, flow, and pipe sizes, we calculate P1, P2, Q, D, etc. When we have a trunk line and short laterals connecting to the trunk lines, yielding increasing flow rates as you go towards the delivery point. Calculating backwards, knowing the required minimum delivery pressure to find P at various tie-in points (nodes) backwards to the original source.

When we calculate the Pressures required to move the gas in theory, what is the way to ensure the model or calculation results are realized as much as possible, is it done with pressure control valves at the tie in points to hold back pressure and ensure that the PCV is set at the calculated pressure or more...?

e.g. deliver point is 800PSIG, working back, we have 3 tie in points at 830psig, 850psig, 900psig, and a beginning source at 950psig. obviously, the flow may not be continuous, but assuming that there is process or flow upset that causes one of the pressures to be much higher, it is it fair to assume that the upstream tie-in point may not be able to flow easily? hence, my question on how to control the pressure? I have not have much pipeline experience.

Thanks in advance.

As much as possible, do it right the first time...

 

[zdas04]

Why exactly do you care if the " the model or calculation results are realized as much as possible". And why on earth would you spend money to force a tie in point to enter the system at 830 psig when the trunk at that point is actually 750 psig? It makes a lot more sense to put discharge control valves on that rare piece of equipment that requires a minimum discharge pressure (e.g., things like screw compressors that are relying on dP across the skid to move lubricating oil, a rare practice, but not unheard of). For flow streams that do not require a minimum discharge (most flow streams), just let it float on line pressure. Your model is a tool to help you understand possible outcomes, it is not the ruler of your destiny.

[bold]David Simpson, PE[/bold]
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. Galileo Galilei, Italian Physicist

 

[LittleInch]

You need to work out how flow in those laterals are controlled.

On pressure, on flow or some other means.

In a multi entry line it's changes to your flows that will impact on the upstream and downstream pressure.

It's more common to control on flow or mass and let the pressures take care of themselves.

If you really want to control on pressure then you can, but this is a system design issue. flow and pressure are intertwined so you need to work out what the limits of your system and entry lines are

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

A birds eye description of the complete pressure / flow control loop at each feeder, starting from its source (a booster compressor in most cases), is the only way to help you with this. Post us a sketch of this together with any pressure control devices / limitations at the final delivery point. Many other factors may also come into play, such as gas quality that may make one lateral's supply preferable over others in case of turndown or operational upsets, startup scenarios, etc.

 

[engr2GW]

Thanks everyone; I was doing the calculation to have an idea, knowing what gas rate is expected at each tie in point, and knowing that the delivery pressure is at the end point. Than I can figure how to control the flow at each point (mostly at compressor discharge at each point). And also check the dp and velocity in each segment to ensure it is within recommended limits.

georgeverghese, per your request, I have attached a sketch with some numbers of the calculations, any advice will be helpful. Note:
1. Segment EF is the last segment with F being the last point
2. Flow in each segment is showing what is flowing from first to second point before the flow at the second point is added to show total flow in the next segment.



As much as possible, do it right the first time...

 

[georgeverghese]

A cross check with isothermal compressible flow equation ( Weymouth , Panhandle is a little antiquated for me) yields a backpressure of 965psig at A, which is lower than your result of 985psig. The discrepancy is largest at segment EF (the other segments are low dp / low velocity) where the pressure at E is 960psig per my calcs compared with your result of 978psig. I've used a roughness factor of 0.046mm for corroded CS piping. I've used the same z values as in your calcs.
I've also done a verification check of my results using incompressible flow routine ( since dp in EF is less than 10% of feed pressure and line velocity is only 8m/sec), and the result is the same as with compressible flow.

No process controls sketch/narrative provided in your response. You may need to get plant owner's Operations team feedback on this also.

 

[engr2GW]

Thank you very much, georgeverghese.
A couple of quick points:
1. Each nodes (tie in points) are discharge of compressors on site tie-ing into the PL. the compressors have suction control valves from the production equipment, and a control valve at the discharge side before PL.
2. I have attached an expanded network proposed, delivery point J (DSB suction compressor station) is 800psig.
a. the DP/100' in some of the segments are greater than 1 due to pipe size and flow rate, is this typically a big concern, other than compression cost?
b. the velocities are okay, unless for a few that are very low, I'm assuming the low fps will be a problem if we have liquid drop out that could be corrosive to the pipe?

based on the attached, I am considering a loop to take the "20 new wells" node directly to the DSB or tied into the 10" segment IJ somewhere, and maybe split some flows from G, H, and I to join the new line to carry the "20 new wells).
Thanks again for your input in advance, the previous post by everyone got me digging more.

As much as possible, do it right the first time...

 

[georgeverghese]

The flows in the 2nd post are much more than in the first post, I've not checked the results in this second post, but it is clear to me the backpressures predicted in your results with Weymouth or Panhandle are higher, so that may result in erroneous conclusions about what the max backpressures at A or any other feeder lateral point of concern may be. See if you can run these dp calcs with some other program or calc routine to get a comparison check.

The line sizes selected will depend a lot on total compression + pipeline CAPEX + compression OPEX over the expected facility life. As long as velocities are less than 60fps, corrosion inhibitor coating on the pipewalls wont erode away.

Check that the setting of the discharge high set PIC on each of the feeder compressors is somewhat higher than the max backpressure predicted. This high setPIC should reset compressor speed to reduce throughput once delivery pressure exceeds PIC setpoint. Obviously, there should be a throttle PIC / PCV on the DSB compressor suction to handle high feed pressure during pipeline restart.

 

[georgeverghese]

At most gas production stations, there is some kind of gas treatment : gas dehydration and acid gas removal or hydrocarbon dewpoint control unit at the least. If this were true in your case, a low set backpressure PIC / PCV would be required for operating cases when pipeline pressure is less than normal operating pressure on the TEG / acid gas removal column. This would throttle the exit into the pipeline at low pipeline pressures so as to maintain the min required pressure for the columns / other gas treatment units to operate. At higher pressures, this PCV would be wide open.

 

[engr2GW]

@georgeverghese, you are correct. the DSP station has those treating capabilities and the precise pressure controls. But not all of the upstream gathering systems has that, actually, most of them are just pulling from the production facilities and headed to the treating facilities.
For now, I do not have an alternate modelling tool/software, I just (for now) compare between different available flow equations that suit the conditions more.

As much as possible, do it right the first time...